Scientific Publications - Work Done by Microbiology Reader Bioscreen C

World Journal of Microbiology and Biotechnology (2001), 17(6), 615-625

Antibacterial activity associated with  Lactobacillus gasseri ATCC 9857  from the human female genitourinary tract

Charteris, William P.; Kelly, Phillip M.; Morelli, Lorenzo; Collins, J. Kevin

ABSTRACT

The 10-fold concentrated spent MRS culture cell-free supernatant concentrate [(cCFS)] of the human female genitourinary tract isolate Lactobacillus gasseri ATCC 9857 was shown to exhibit antibacterial activity towards gram-positive sporogenous and asporogenous fermentative eubacteria in liquid and on solid media under conditions that eliminated the activity of lactic acid (fl-glycerophosphate) and hydrogen peroxide (catalase). The antibacterial activity of the cCFS was characterized by automated turbidometry (Bioscreen^') and non-linear regression analysis (Gompertz model) using MRS broth cultures of the indicator strain L. acidophilus ATCC 11975. It exhibited a bactericidal mode of action, sensitivity to trypsin and proteinase K, partial sensitivity to pepsin and pronase E, partial heat stability at 121 °C for 15 min, and retained significantly more activity following exposure to pH 3.0 and 5.0 compared with pH 7.2 and 9.0. The inhibitory spectrum included a wide range of Lactobacillus species, Bidobacterium bidum, B. infantis and B. catenulatum, Lactococcus cremoris, Leuconostoc cremoris, Pediococcus pentosaceus, Bacillus cereus, Clostridium tyrobutyricum, C. pasteurianum, C. sporogenes, Staphylococcus carnosus, and Enterococcus faecalis. Although partial inhibition of Escherichia coli ATCC 25922 by cCFS was observed in liquid medium, inhibition of freshly isolated human uropathogenic E. coli strains could not be demonstrated on TSB agar plates by agar well diffusion. Following partial resolution by gel permeation FPLC on Superose-12, the fractionated cCFS was shown to comprise at least two inhibitory peptides (3.05 and 5.27 kDa) as well as aggregated inhibitory peptide material (21.65, 41.50, 81.20, and 120.90 kDa). The 3.05 kDa peptide, designated Gassericin D, inhibited L. acidophilus strains ATCC 11975 and ACA-DC 241. The 5.27 kDa peptide, designated Gassericin C, inhibited L. gasseri strain UCSC LF221 Snb and En. faecalis DPC 3319. The aggregated 21.65 kDa peptide material strongly inhibited L. acidophilus ATCC 11975 and weakly inhibited Listeria inocua DPC 3306. The aggregated 41.50 kDa peptide material strongly inhibited Ba. cereus DPC 3316 and weakly inhibited L. acidophilus ACA-DC 241. The ability of L. gasseri ATCC 9857 to produce bacterioci-like activity may be of importance in the biopreservation of nutraceuticals and in the management of female genitourinary and gastrointestinal tract infections involving En. faecalis.

Introduction

Lactobacillus species enjoy a time-honoured reputation as health promoters especially in the healthy female genitourinary tract (Reid et al. 1987; McGroarty 1993; Charteris et al. 1997; Barbes & Boris 1999). They adhere to uroepithelial and vaginal cells, and constitute the predominant members of the indigenous genitourinary microbiota. Lactobacilli are assumed to contribute to the maintenance of a healthy ecological balance and colonization resistance to uropathogens. Moreover, intravaginal instillation, oral administration, and intra­muscular injection of selected lactobacilli have been shown to be useful for the management (prevention and treatment) of genitourinary tract infection.

The production of inhibitory metabolites, such as hydrogen peroxide, biosurfactants, and bacteriocin-like substances, together with an ability to adhere to the female genitourinary mucosa, to co-aggregate with uropathogens and promote their elimination, and to resist the antimicrobial activities of microbiocides, including spermicides, and antibiotics, are among the mechanisms by which lactobacilli are thought to contrib­ute to genitourinary health. These properties constitute important strain selection criteria for bacteriotherapeu­tic lactobacilli in the management of recurrent genitour­inary tract infection.

Table 1. Screening of candidate indicator strains in the EU FLAIR Project culture collection.

ND - Not determined.

a Strains were made available as part of EU FLAIR Project CT91-0053 by the following:

1. Amercian Type Culture Collection, Cockeysville, Maryland, USA; 2. Prof George Kalantzopoulos, Agricultural University of Athens, Athens, Grrece; 3. Prof Range Fonden, Panova Partner AB, Aarla Group, Stockholm, Sweden; 4. Dr Anton Weerkamp, Netherlands Institute for Dairy Research, Ede, The Netherlands; 5. Dr Brian Phillips, National Collection for Food Bacteria, Reading, United Kingdom; 6. Prof Lorenzo Morelli, Instituto di Microbiologia, Catholic University of Piacenza, Italy.

^b Susceptibility to L. gasseri ATCC 9857 cCFS determined in liquid medium by automated turbidometry.

Bacteriocinogenic activity in urine by Lactobacillus rhamnosus GR-1, an indigenous distal urethral isolate, has been reported (McGroarty & Reid 1988). The antimicrobial was bactericidal towards a pyelonephri­togenic Escherichia coli, non-dialysable, heat labile, lipophilic, and of low molecular weight (greater than 12,000-14,000 Da). Moreover, Ocana et al. (1999) have reported the production of a heat-resistant bacteriocin­like substance by a vaginal isolate of L. salivarius subsp. salivarius. In this study, we report the partial charac­terization of a bacteriocin-like substance produced by the female genitourinary isolate L. gasseri ATCC 9857 (also known as Doderlein's bacillus, L. acidophilus NCTC 2948, and L. acidophilus VPI 11089) (Fujisawa et al. 1992). Some evidence of its antimicrobial activity has been reported previously (Vincent et al. 1959; Barefoot & Klaenhammer 1983; Lewus & Montville 1991).

The cell-free supernatant concentrate (cCFS) of L. gasseri strain ATCC 9857 (together with those from the neonatal faecal isolates L. gasseri strains UCSC L32012BR and UCSC LF221 Snb) was originally shown to inhibit the growth of some closely related lactic acid bacteria and bifidobacteria during a preliminary screen­ing programme to identify antibacterial activity among strains deposited in the EU FLAIR Project CT91-0053 Culture Collection (Charteris & Kelly 1993a, b). A preliminary partial characterization of the antibacterial activity is now reported which provides confirmatory evidence of bacteriocinogenic activity in L. gasseri (Toba et al. 1991; Kawai et al. 1994; Tahara et al. 1996, 1997) and suggestive evidence that it may be active in host defence in the female genitourinary tract.

Materials and methods Strains and culture conditions

All strains were maintained at -20 °C on glass beads in broth containing glycerol (40% v/v) (Jones et al. 1984) and serially transferred anaerobically at 37 °C (aero­bically for Bacillus cereus) for 48 h three times prior to assay. Three groups of potential indicator strains were used to determine the antibacterial spectrum of L. gasseri ATCC 9857. The first comprised Lactobacillus and Bidobacterium strains from the EU FLAIR RTD Shared-cost Project CT91-0053 Culture Collection (Table 1). The second comprised strains commonly used during the EU ECLAIR RTD Programme for bacterio­cin screening (Table 2). The third comprised E. coli ATCC 25922 and six human uropathogenic E. coli strains that had been freshly isolated on MacConkey agar from the urine of patients presenting with acute pyle­onephritis at Cork University Hospital, Ireland. Growth media were from Oxoid (Unipath Ltd., Basingstoke, UK) and included: MRS for lactobacilli (De Man et al. 1960); TPY for bifidobacteria; BHI for bacilli, entero­cocci, listeria and staphylococci; RCM for clostridia; and TSB (soybean casein digest medium USP) for E. coli. Soft agar overlay was prepared with 0.7% w/w agar technical no. 1 (Oxoid) and buffered with 1.7% w/w fl-glycero­phosphate (Sigma Chemical Co. Ltd., Poole, UK).

(Bio) Chemicals

All chemicals, except chloroform (BDH Lab. Supplies Ltd., Poole, Dorset, UK), were from Sigma. Stock solutions of catalase (EC 1.11.1.6), trypsin (EC 3.4.21.4), pronase E (EC 3.4.24.31), pepsin (EC 3.4.23.1), and proteinase K (EC 3.4.21.64) were prepared in distilled water (0.5 mg/ml), Tris-HCl pH 8.0 (0.2 mg/ ml), Tris-HCl pH 8.0 (0.25 mg/ml), sodium citrate pH 3.0 (0.2 mg/ml), and sodium phosphate pH 7.2 (0.1 mg/ ml), respectively.

Table 2. Screening of EU ECLAIR Programme candidate indicator strains.

Phenotypic characterization of strains

The taxonomic position of L. gasseri strain ATCC 9857 (together with L. gasseri strains UCSC L32012BR, LF221 Snb, and ATCC 4962) was assessed using a combination of 49 carbohydrate acidification tests (API 50 CHL, BioMerieux SA, La Balme les Grottes, 38390 Montalieu Vercieu, France) and 19 enzymatic tests (API-ZYM, BioMerieux). Carbohydrate acidification was scored after 48 and 72 h incubation at 37 °C and the profile status determined using the APILAB Plus software (V4.0 database). Enzymatic tests were per­formed at 37 °C for 4 h and scored according to manufacturer's instructions.

Preparation of pH-neutralized concentrated cell free supernatant (cCFS)

Lactobacillus gasseri ATCC 9857 was grown to late­logarithmic phase (48 h) in MRS broth at 37 °C under anaerobic conditions (BBL GasPak^®, Becton Dickinson Inc., New Jersey, USA). CFS was obtained by centri­fugation at 9000 x g for 30 min using a Sorvall^® RC-5B Plus centrifuge (Kendro Lab. Products Ltd., Herts., UK), freeze-dried using liquid N2 and a Hetosicc Type CD52 drier (Heto Lab. Equipment A/S, Copenhagen, Denmark), reconstituted to one-tenth original volume in 0.1 M sodium phosphate buffer pH 6.5 containing 0.05 mg catalase/ml, and stored at -20 °C.

Detection and confirmation of antibacterial activity

Antibacterial activity was determined spectrophotomet­rically (wide band filter, 620 nm) using the Bioscreen^TM automated turbidometer (LabSystems Oy, Pulttitie 8, 00880 Helsinki, Finland) (Heinon et al. 1987). The analyser was used to monitor bacterial growth kinetically during the incubation period. Up to 200 samples were run in parallel, thereby enabling multivar­iate studies to be performed. Pre-warmed multi-well plates of 400 µl capacity were charged with 40 µl of an active overnight culture of candidate indicator strain, 40 µl of cCFS, and 320 µl of growth medium appropriate to the candidate indicator strain to be assayed. Optical measurements were recorded under aerobic growth conditions during incubation at 37 °C for up to 76 h.

Following conversion of Bioscreen^TM data source files (*.res) to Microsoft Excel^® file format (*.xls), optical density data were log-transformed and analysed by non­linear regression using the Marquardt `least squares'

iterative parameter estimation algorithm (Marquardt 1963) to determine the parameters of the Gompertz equation (Whiting & Buchanan 1997):

Y_t = Yo + C exp{- exp[-B(Time - M)] }

where Yo and Y_t are the loge transformed population sizes at two and time t, respectively, C is the change in cell density between tzero and the start of the stationary phase, B is the relative growth rate, and M is the time when maximum growth rate is achieved.

Initial parameter estimates were derived from X-Y plots of the dependent variable (loge OD620) vs. the independent variable (time). The final parameter esti­mation process terminated successfully after <8 itera­tions in all cases. In turn, these parameters were used to calculate the magnitude of the impact of cCFS addition on the lag time (h), M-1 /C, and specific growth rate, BC/exp(1), of the indicator strain under study. Results are expressed as `difference from control'. In view of the fact that the OD data sets were not used for predicting viable counts but for comparative purposes only (test vs. control), no attempt at calibration was made to relate OD to viable counts.

Antibacterial activity detected in liquid media using the Bioscreen^TM was confirmed as bacteriocin-like using two agar plate methods: well diffusion and deferred antagonism. In the former method, petri plates (9 cm) containing 15 ml of solid agar medium were overlaid with 4 ml soft agar overlay (containing 1.7% w/w fl­glycerophosphate) to which had been added 400 µl of an active indicator strain and 40 µl of catalase stock solution (0.5 mg/ml). After solidification and air drying, 5 mm wells were aseptically cut in the medium with a cork borer (Sigma) and charged with 40 µl of cCFS. Following pre-diffusion at room temperature for 1 h, plates were incubated anaerobically at 37 °C (aero­bically for Ba. cereus) for 24 h, and inhibition zone diameters measured using a sliding calliper. In the latter method, L. gasseri ATCC 9857 was spotted onto MRS agar and grown anaerobically at 37 °C for 48 h. Thereafter, surface colonies were exposed to chloroform vapour for 1-2 min, and overlaid with 4 ml soft agar overlay as described above. Petri plates were incubated anaerobically at 37 °C (aerobically for Ba. cereus) for 24 h, and inhibition zone diameters measured.

Sensitivity to proteases, pH, and heat

The Bioscreen^TM was used to monitor the effect of protease activity, pH exposure, and heat treatment on cCFS antibacterial activity. For protease sensitivity, each microtitre well assay comprised 40 µl of L. gasseri ATCC 9857 cCFS, 40 µl of buffered protease solution, 40 µl of an overnight culture of indicator strain, and 280 µl of growth medium appropriate to the indicator strain. For pH exposure, the buffer solutions free of protease were used as described above. For heat sensitivity, the cCFS was heated at either 60 °C for 30 min or 121 °C for 15 min prior to assay of antibac­terial activity as described above. Growth of the indicator strains in the absence of protease addition, pH exposure, and heat treatment served as control.

Gel filtration

Gel filtration was performed at room temperature with u.v. detection (280 nm) by FPLC equipped with Superose 12 (12 x 30 mm) (Pharmacia AS, Uppsala, Sweden). Fifty microlitre of a 1:4 dilution of L. gasseri ATCC 9857 cCFS were applied to the column equili­briated with 0.1 M Tris-HCl buffer (pH 6.5) and eluted with the same buffer for 1 h at a flow rate of 0.5 ml/min. Molecular weight determination of activity peaks was performed using a gel filtration calibration kit (Phar­macia) and a mixed linear second-order polynomial estimation model. The Bioscreen^' was used to monitor antibacterial activity in selected fractions towards L. ac­idophilus ATCC 11975, L. gasseri UCSC LF221 Snb, L. acidophilus ACA-DC 241, En. faecalis DPC 3319, Ba. cereus DPC 3313, and Listeria inocua DPC 3306. Each microtitre well assay comprised 40 µl of S-12 fraction, 40 µl of an overnight culture of indicator strain, and 320 µl of growth medium appropriate to the indicator strain. Growth of the indicator strain in the presence of 0.1 M Tri s-HCl pH 6.5 (absence of S-12 fraction) served as control. The magnitude of inhibition at mid-log phase of growth (end-log for Ba. cereus only) was determined.

Reproducibility

 The results presented herein were observed for a single L. gasseri ATCC 9857 cCFS preparation and are repre­sentative of a larger dataset in which each experimental treatment was repeated at least three times with each cCFS preparation and each experiment was replicated at least three times with freshly prepared cCFS preparations.

Results

Preliminary experimentation

Lactobacillus gasseri strains ATCC 9857, UCSC L32012BR, and UCSC LF221 Snb were previously shown to possess two bacteriocin-like activities when tested against 46 Lactobacillus and Bifidobacterium isolates (Charteris & Kelly 1993a; Charteris et al. 1994). The antimicrobial spectrum of each cCFS ap­peared to be strain-specific. In this regard, the inhibitory spectrum of strain ATCC 9857 exclusively included L. acidophilus ATCC 314, L. rhamnosus NCFB 86, and L. casei UCSC 3026, while those of strains UCSC L32012BR and UCSC LF221 Sn exclusively included L. rhamnosus strains NCFB 330 and 1858 and L. casei ACA-DC 212.2 and 212.3, respectively. The bacteriocin­like activity of L. gasseri ATCC 9857 was selected for further study, the results of which are reported herein.

Taxonomic position

The efficiency of the API 50 CHL commercial kit for identification of L. gasseri was evaluated using strains ATCC 9857, UCSC L32012BR, UCSC LF221 Snb, and ATCC 4962 that had previously been identified by DNA-DNA hybridization. All strains were misidenti­fied as L. acidophilus (data not shown).

Inhibitory spectrum

The cCFS of L. gasseri ATCC 9857 inhibited at least one strain from almost all the Lactobacillus and Bido­bacterium species tested (Tables 1 and 2). Lactobacillus acidophilus group strains and B. bidum strains were particularly susceptible, whereas L. caseilrhamnosus group strains were particularly insensitive. In view of its relatively high sensitivity, L. acidophilus ATCC 11975 was selected as the preferred indicator strain for further characterization of the bacteriocin-like activity.

The antibacterial activity of L. gasseri ATCC 9857 cCFS towards various gram-positive species was tested in liquid (Bioscreen^TM) and on solid media. Inhibitory activity was found not to be limited to lactobacilli and bifidobacteria but extended over a wide range of aspor­ogenous and sporogenous fermentative eubacteria. It was evident that determination of inhibitory activity by deferred antagonism (colony agar overlay following chloroform treatment) was less sensitive than by agar well diffusion (following freeze-dry concentration of CFS) for Ba. cereus, C. tyrobutyricum, and Staphylococ­cus carnosus, which may suggest a possible chloroform susceptibility effect on antibacterial activity. However, it was noted that inhibitory activity towards L. helveticus DPC 3321 was lost on freeze-dry concentration. More­over, it was also evident that determination of growth inhibition using the Bioscreen^TM method appeared to be as good as that by agar well diffusion with inhibitory activity towards strain DPC 3321 evident.

Although partial inhibition of E. coli ATCC 25922 was observed in liquid medium (Figure 1), it could not be confirmed on solid media. Moreover, addition of cCFS to E. coli ATCC 25922 during mid- and late-logarithmic growth phases did not impact on growth (data not shown). Furthermore, inhibitory activity towards freshly isolated human uropathogenic E. coli strains was not observed on solid media either by deferred antagonism or agar well diffusion (data not shown).

Mode of action

The effect of the addition of cCFS to L. acidophilus ATCC 11975 during mid-logarithmic growth was used to establish a bactericidal mode of action. An immediate and sustained (for up to 3 h) fall in OD620 occurred upon addition. Moreover, the bacterial count was >3 log cycles lower 3 h after addition than in the control (data not shown).

Sensitivity to proteases, pH, and heat

To establish the proteinaceous nature of the antibacte­rial activity, the effect of protease addition on cCFS-inhibitory activity towards L. acidophilus ATCC 11975 was examined (Table 3). Proteinase K and trypsin completely inactivated inhibitory activity, whereas pep­sin and pronase E caused only partial inactivation.

To determine the level of antibacterial activity under genitourinary and lower gastrointestinal pH conditions, cCFS was exposed to buffers with pH values ranging from 3.0 to 5.0 and 7.2 to 9.0, respectively, and antibacterial activity towards L. acidophilus ATCC 11975 during the early-logarithmic growth phase determined spectropho­tometrically (Table 3). Although cCFS retained signifi­cant antibacterial activity at pH 3.0 and 5.0, considerable loss of activity was observed at pH 7.2 and 9.0.

Heat sensitivity was also determined spectrophoto­metrically after heat treatment of cCFS and when L. acidophilus ATCC 11975 was in the early-logarithmic phase of growth (Table 3). Little loss of inhibitory activity occurred following heat treatment at 60 °C for 30 min and 121 °C for 15 min.

Molecular weight determination of inhibitory fractions

Gel permeation FPLC on Superose-12 resolved two inhibitory peptides (3.05 and 5.27 kDa) and aggregated peptide inhibitory material (21.65, 41.50, 81.20, and 120.90 kDa) (Figure 2a, b). The 3.05 kDa peptide, designated Gassericin D, inhibited L. acidophilus strains ATCC 11975 and ACA-DC 241. The 5.27 kDa peptide, designated Gassericin C, inhibited L. gasseri strain UCSC LF221 Snb and E. faecalis DPC 3319. The aggregated 21.65 kDa peptide material strongly inhibit­ed L. acidophilus ATCC 11975 and weakly inhibited Li. inocua DPC 3306. The aggregated 41.50 kDa peptide material strongly inhibited Ba. cereus DPC 3316 and weakly inhibited L. acidophilus ACA-DC 241.

Discussion

Lactobacilli were first described as normal constituents of the microbiota of the female genitourinary tract in 1892 when 136derlein discovered the vaginal lactobacil­lus in his study of puerperal sepsis (Charteris et al. 1997). The number of lactobacilli ranges from 1 x 10¹ to 5 x 10⁹ c.f.u./ml of vaginal fluid from healthy women, with a mean viable count of about 10⁸ c.f.u./ml. The clinical significance of lactobacilli was demonstrated in patients with vaginitis with the absence of lactobacilli being correlated with increasing severity of infection. Lactobacilli predominate on the mucosal surfaces of the female genitourinary tract where they thrive at the low pH, which is generated by acid epithelial cell metabolites (Redondo-Lopez et al. 1990).

a Calculated from Gompertz model parameter estimates (average of five determinations) following non-linear regression analysis as described in Materials and methods.

Figure 1. Growth inhibition E. coli ATCC 25922 L. gassed ATCC 9857 cCFS.

Table 3. Effect of protease addition, pH exposure, and heat treatment on growth inhibition of L. acidophilus ATCC 11975 by L. gasseri ATCC 9857 cCFS.